Method and device for detecting local defects on a reflective surface

Description

[0001] The invention relates to a method and a device for detecting local defects on a reflective surface, wherein the device can generate at least one pattern for reflection on the reflective surface and comprises at least one camera and a data processing unit. The invention further relates to a computer program product and a computer-readable data carrier. Methods in which a pattern known in shape and position is reflected in a reflective surface and the reflected image is viewed and evaluated with a camera are known for measuring the shape of reflective surfaces. The term deflectometry is also used for such methods.

[0002] Such deflectometry methods already exist in a variety of forms. In known methods, a camera is used to look at the reflection of a pattern mirrored on the surface. This pattern is encoded in some way, either spatially or temporally, as a sequence of different patterns that are captured sequentially, resulting in one or more corresponding images. These patterns can then be identified in the one or more images. Thus, when using a digital camera, each camera pixel can be assigned a point on the pattern(s) that it is looking at. Through prior calibration, the system knows the spatial location of the camera and the pattern(s). With this information, various geometric properties of the surface can then be determined, such as inclination, curvature, and, depending on the design of the known system, also the shape of the surface.These surface properties can then be used to determine whether the surface has defects. Examples of such a procedure are presented in documents WO 2007 / 115621 A2 and US 2019 / 0287237 A1.

[0003] Document DE 10 2018 118 602 B3 discloses a method for detecting and analyzing surface defects of three-dimensional objects with a reflective surface, in particular motor vehicle bodies, as well as a device for this purpose. In these methods, the surface defects are identified by evaluating at least one image, recorded by at least one camera in the form of a raster graphic of pixels, of an illumination pattern projected onto at least a part of the surface by at least one first illumination device, using a two-dimensional raster coordinate system. The disclosure relates in particular to the identification of surface defects based on two-dimensional image information using image processing algorithms, for example, by constructing a function graph for a group of pixels identified as belonging together.From the function graph, the first derivative of the found functional relationship can be determined, and surface errors can be identified based on characteristic changes and / or the loss of the first derivative of a found functional relationship.

[0004] However, these methods are comparatively complex. Extensive algorithms must be calculated to decode the patterns and reconstruct the measured surface. This significantly reduces the processing speed. Furthermore, many of these methods require multiple images to resolve the coding, or they necessitate large-area, uniquely surface-coded patterns, which are relatively difficult to produce. Calibrating these systems is also complex, and in many cases, large plane mirrors are required. This is particularly challenging for large or highly curved surfaces.

[0005] The object of the present invention is therefore to specify a method and to create a device that achieves a higher speed in the detection of defects on reflective surfaces.

[0006] The above problem is solved by a method having the features of claim 1, a device having the features of claim 6, a computer program product having the features of claim 11, and a computer-readable data carrier having the features of claim 12.

[0007] In particular, the problem is solved by a method for detecting local defects on a reflective surface using a device that can generate at least one pattern for reflection on the reflective surface and comprises at least one camera and a data processing unit. The pattern has at least one, essentially linear, light-dark transition, wherein the positioning and orientation of the camera are known, the camera captures the pattern reflected on the surface and generates image data of the reflected pattern, which are transmitted from the camera to the data processing unit.The data processing unit detects local defects on the reflective surface based on an evaluation of at least one property of the at least one light-dark transition in the image data of the reflected two-dimensional pattern, wherein the at least one property comprises a contrast in a predefined area of ​​the at least one light-dark transition in the image data. No geometric measurement of the surface is performed; rather, only local deviations from the ideal shape or local defects are detected. These can be large-area or small-area. The high inspection speed achievable with the method according to the invention is advantageous, since the calculations are less complex and therefore not as costly as in the prior art.

[0008] Within the scope of this invention, the reflective surface is not only a completely reflective surface, but also, in particular, a partially transparent surface of a solid or liquid object, for example, a glass pane. Therefore, the invention as a whole is also particularly suitable for detecting defects in the surface of windshields for motor vehicles or other curved panes. The reflective surface reflects electromagnetic radiation from the visible wavelength range at least partially, for example, over a wavelength range from 380 nm to 780 nm or a portion thereof. Additionally, the surface can also reflect electromagnetic radiation in the infrared range (wavelength greater than 780 nm) and / or ultraviolet radiation (wavelength less than 380 nm).

[0009] In one embodiment, the defects of several reflective surfaces can be determined, e.g., the top and bottom of a glass pane and / or several superimposed glass panes or other transparent objects.

[0010] The camera is a digital camera, for example a matrix or line scan camera, whose position and orientation in space, particularly with respect to the reflective surface, are known. The camera's position and orientation can be determined through calibration. For example, by knowing the viewing direction for each pixel of the camera, the corresponding pattern position can be determined with good spatial resolution and simple optical means from the position of the pattern area reflected on that pixel. This enables a fast and accurate local mapping of a detected defect to its position on the surface. The camera determines a brightness value and / or a color value for each image of the pattern and each pixel, representing the image data that makes up an image of the pattern. This image data is then forwarded to the data processing unit and analyzed there.

[0011] The data processing unit can be integrated into the camera as a module or form a separate unit. In the latter case, image data is transmitted from the camera to the data processing unit either via cable or wirelessly. The data processing unit comprises a processor, which is a functional module that interprets and executes instructions / commands of an algorithm, as well as an instruction control unit, an arithmetic unit, and a logic unit. The processor can include at least a microprocessor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA – a digital integrated circuit into which a logic circuit can be programmed), a discrete logic circuit, and any combination of these components. The data processing unit can also include a memory module, an input module (e.g., a USB flash drive), a USB flash drive, ...The data processing unit (DPU) may consist of a keyboard or touchpad, a power supply module (e.g., battery), and a display module (e.g., screen). The DPU can be a physical hardware resource, such as a smartphone, desktop computer, server, notebook, cluster / warehouse scale computer, embedded system, or similar, or a virtualized computer resource. Furthermore, the DPU may include a transceiver for exchanging data with the camera.

[0012] The pattern can be displayed on a screen and positioned above the reflective surface using a pattern holder. The camera observes and records the reflected image of the pattern. The camera is also positioned above the reflective surface, i.e., on the same side of the reflective surface as the pattern. The pattern is positioned so that it faces the reflective surface.

[0013] To further increase accuracy, in one embodiment the pattern can be displayed on one or more displays. This allows the pattern's design to be adapted to the characteristics of the reflective surface (e.g., its extent). In one embodiment, the pattern can be black and white or color. It contains at least one essentially linear light-dark transition, which forms a continuous line that appears light or dark depending on the viewing angle. The line can be completely straight or curved, or partially straight and / or partially curved. Viewed perpendicular to the line, the light-dark transition forms an area of ​​high brightness (light area) followed by an area of ​​low brightness (dark area), or vice versa, with the line between the light and dark areas, i.e.,This is created by the transition from light to dark or dark to light. In one embodiment, this transition is sharply defined, meaning the gradient of brightness across the transition is comparatively large; for example, the relative change in brightness per mm is at least 60% per mm. Here, brightness (value) refers to the respective intensity of the light within a given wavelength range.

[0014] According to the invention, the method comprises a property of the at least one light-dark transition in the image data in the form of a contrast in a predetermined area of ​​the at least one light-dark transition. In an exemplary embodiment, the method can further comprise a shape of the at least one light-dark transition and / or a change in brightness over a predetermined distance across the at least one light-dark transition, each in the image data. This is based on the understanding that a defect on the reflective surface results in a deviation in the image data in the shape of the light-dark transition line and / or in the contrast of an area of ​​the line or in the change in brightness of the light-dark transition, which is detected during the analysis of the image data in the data processing unit.Contrast, in this context, refers to a value calculated from a minimum brightness value (Imin) within a given area of ​​the image data and a maximum brightness value (Imax). The resulting contrast value is compared to a predefined contrast target value or range, which is determined, for example, by the camera parameters. If the calculated contrast value differs from the contrast target value or lies outside the contrast target range, an error is detected. Regarding the shape of the line, a comparison of the determined line shape with a predefined line shape (so-called pattern matching) is performed. For this purpose, the data processing unit first identifies the line in the image as the brightest line segment (in the case of a light line) or the darkest line segment (in the case of a dark line), determines its shape, and detects the adjacent light and dark areas.A defect is also detected if the line shape deviates from a predefined tolerance range. For example, the surface may exhibit a change in direction, inclination, or curvature, which is detected as a deviation in the line shape relative to the pattern. The system also determines whether the measured brightness change (i.e., brightness gradient) over a predefined distance of the light-dark transition deviates from a target brightness value or range. If so, a surface defect is detected. For instance, the light-dark transition in the image may appear less sharp and more blurred in the area of ​​a defect than in the original pattern, which can be represented by the brightness gradient. The gradient in the camera image is therefore smaller than in the original pattern.

[0015] Since the camera's position and viewing direction are known, and thus information is available about which area of ​​the image data captures which local area of ​​the reflective surface, when a defect is detected, for example, using one or more of the analyses mentioned above, the detected defect can be assigned to a specific position on the surface. The dimensions of the defect can also be determined by having the data processing unit ascertain over which area of ​​the surface the deviations from the target values ​​or target value ranges, or from the line shape, extend.

[0016] In one embodiment, the position of the pattern relative to the surface is unknown, i.e., not defined. This means that calibration of the pattern to the camera and the reflective surface is unnecessary. The exact position of the pattern is not required for the present task, since the goal is not to measure the reflective surface itself, but merely to determine the position and / or dimensions of a defect on the surface. A defect is defined here as a deviation from the desired surface finish and / or structure that impairs the optical properties of the surface, particularly its properties in the reflection of electromagnetic radiation in the visible wavelength range. Examples of defects include inclusions in the surface area, surface irregularities, and the like.To correctly evaluate the properties of the light-dark transition, the camera is focused on a medium distance from the reflective surface. Furthermore, the camera aperture is set so that the line of the at least one light-dark transition is imaged in such a way that the light-dark transition is detectable.

[0017] In one embodiment, the pattern has at least one stripe forming two light-dark transitions, and the data processing unit evaluates the width of this stripe to detect local defects on the surface. The stripe can be a light stripe (light stripe) relative to a dark environment or a dark stripe relative to a light environment in the pattern. In another embodiment, the pattern can be a striped grid. The stripe serves only to generate two light-dark transitions but does not represent any encoding with respect to the pattern.

[0018] In one embodiment, the pattern and / or the surface undergo relative movement to each other during the camera's recording of the reflected pattern. This allows for the measurement of a large reflective surface compared to the size of the pattern.

[0019] The problem is further solved by a device for detecting local defects on a reflective surface, which can generate at least one pattern for reflection on the reflective surface and comprises at least one camera and a data processing unit, wherein the pattern has at least one substantially linear light-dark transition, wherein the positioning and orientation of the camera are known, wherein the camera is configured to record the pattern reflected on the surface and to generate image data of the reflected pattern and to transmit it to the data processing unit, wherein the data processing unit is configured to determine local defects on the surface based on an evaluation of at least one property of the at least one light-dark transition in the image data of the reflected pattern.wherein the at least one property comprises a contrast in a predefined range of the at least one light-dark transition in the image data. The device enables fast and inexpensive localization of defects on a reflective surface.

[0020] The exemplary embodiments of the model given above are also relevant for the device described above. Reference is made to the explanations above regarding this.

[0021] In one embodiment of the device, the pattern has at least one stripe forming two essentially linear light-dark transitions, and the data processing unit is configured to evaluate the width of the at least one stripe to detect local defects on the surface. As explained above, the use of light stripes or dark stripes and the evaluation of the stripe width in the image data for defect localization represents a very simple embodiment of the present invention.

[0022] In one embodiment of the device, the pattern and / or the surface are arranged such that they perform a relative movement to each other during the recording of the reflected pattern by the camera in order to analyze a large surface.

[0023] The above procedure can be implemented, for example, as a computer program comprising instructions which, when executed, cause a processor of the data processing unit to perform the steps of the above procedure, wherein the computer program includes a combination of the steps and data definitions described above which enable the computer hardware to perform computational or control functions, and / or which represents a syntactic unit which conforms to the rules of a particular programming language and which consists of declarations and statements or instructions required for the functions, tasks, or problem solutions explained above.

[0024] Furthermore, a computer program product is disclosed, comprising instructions which—as described above—cause the above-mentioned device to perform the steps of the above-defined method. Correspondingly, a computer-readable data carrier is disclosed that stores such a computer program product. The computer program product may be a software routine.

[0025] Further advantages, features, and applications of the present invention will also become apparent from the following description of exemplary embodiments and the drawing. All features described and / or illustrated, individually or in any combination, constitute the subject matter of the invention, even independently of their compilation in the claims or their cross-references.

[0026] It shows schematically Fig. 1 shows an embodiment of a device according to the invention for measuring the shape of a reflective surface.

[0027] In Fig. 1The basic setup of a device 10 for detecting defects in a reflective surface 13 of an object 3 to be measured is shown. The device 10 has a camera 1 which observes the pattern 12 of a pattern carrier 2 via the reflective surface 13. In order to determine defects in the reflective surface 13 of the object 3, a coordinate system of the camera 1 is determined using a known camera calibration method. For example, a camera calibration method can be used that observes a planar pattern in at least two different orientations and is described in Zhang, Z., "A flexible new technique for camera calibration", IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 22, no. 11, 2000, pages 1330-1334. Other camera calibration methods are described, for example, in Fraser, Clive S., "Digital camera self-calibration", ISPRS Journal of Photogrammetry and Remote Sensing, no.52(4),1997, pages 149-159 described.

[0028] The digital camera 1 is set up for pixel-wise viewing of the 2-dimensional pattern 12 reflected on the reflective surface 13.

[0029] The pattern 12 can be displayed in various ways. A flat screen, e.g., a TFT screen, is used as the pattern carrier 2, on which any pattern 12 can be displayed. Due to the known pixel dimensions of the flat screen 2, the geometry of the displayed pattern 12 is also precisely known. However, other implementations for a pattern carrier 2 are also conceivable, e.g., interchangeable plates with a measured pattern in a holder. The pattern 12 can, for example, consist of two light stripes with a width B1 and a width B2 on a dark background, separated by a distance greater than the widths B1 and B2 of the two stripes. B1 and B2 can be the same or different. The distance between the two stripes is the distance between two adjacent light-dark transitions of the adjacent stripes.Each stripe has two light-dark transitions: a first light-dark transition from a low-brightness area to a high-brightness area, and a second light-dark transition from a high-brightness area to a low-brightness area.

[0030] In practice, the objects to be examined for surface defects are, for example, a motor vehicle windshield. The reflective surface is the outer surface 13 of the windshield 3.

[0031] Camera 1 captures an image of the stripe pattern and transmits the generated digital data to data processing unit 7.

[0032] The detection of defects on surface 13 is performed by a processor contained in the data processing unit 7, based on the data transmitted by camera 1 and the image of the pattern's light-dark transitions reflected from surface 13. For this purpose, the brightness and / or a color or grayscale value received by data processing unit 7 from camera 1 is evaluated for each pixel of the camera that corresponds to a point in the image. First, the transition lines are located using thresholding operations. If the line is the brightest area in a transition, it is assumed that only the areas where the line is visible appear bright in the image. Accordingly, such a thresholding operation is also possible for a dark line.From the detected intensity maxima, the line and its surroundings are extracted. This surroundings are a curved strip with a certain extent of light / dark areas perpendicular to the line, following the path of the mirrored line. Subsequent calculations, such as determining the local contrast and the local shape and width of the line, are then performed on this strip. It is then determined whether the line's path exhibits any changes compared to the pattern 12, for example, changes in direction. Alternatively or additionally, the width of the stripes extending between adjacent linear light-dark transitions in the areas of high brightness can be determined. If deviations from pattern 12 occur in the light-dark transitions, the data processing unit detects a defect in the surface 13 of the windshield 3.The data processing unit 7 can also assign the defect to a specific position on the surface 13 of the windshield 3, since the camera 1 is calibrated with respect to its position and viewing direction. The extent / dimensions of the defect can also be determined based on the extent of the area in the image captured by the camera 1, by detecting a change in the respective light-dark transition and linking this change to the corresponding positions of the area on the surface 13.

[0033] The steps above are easy to implement and result in high inspection speed. Defect locations and dimensions are determined. The data processing unit can store the determined defects and their dimensions in a memory unit and / or display them on a screen. Reference symbol list

[0034] 1 Camera 2 Pattern carrier, flat screen 3 Windshield 7 Data processing unit 10 Device 12 Pattern 13 Reflective surface of the windshield 3

Claims

1. A method for detecting local defects on a reflective surface (13) by means of a device (10), wherein the device can produce at least one pattern (12) for reflection on the reflective surface and comprises at least one camera (1) and a data processing unit (7), wherein the pattern comprises at least one substantially line-shaped light-dark transition, the positioning and orientation of the camera being known, wherein the camera records the pattern reflected on the surface and generates image data of the reflected pattern which is transmitted from the camera to the data processing unit, wherein the data processing unit determines local defects on the surface based on an evaluation of at least one property of the at least one light-dark transition in the image data of the reflected pattern, characterized in that the at least one property comprises a contrast in a predetermined area of the at least one light-dark transition in the image data.

2. The method according to claim 1, characterized in that further properties of the at least one light-dark transition in the image data comprise a shape of the at least one light-dark transition in the image data and / or a change in brightness over a predetermined distance across the at least one light-dark transition in the image data.

3. The method according to any one of the preceding claims, characterized in that the position of the pattern with respect to the surface is unknown.

4. The method according to any one of the preceding claims, characterized in that the pattern comprises at least one stripe forming two light-dark transitions and the data processing unit evaluates the width of the at least one stripe to determine local defects on the surface.

5. The method according to any one of the preceding claims, characterized in that the pattern and / or the surface perform a relative movement with respect to each other during the recording of the reflected pattern by the camera.

6. A device (10) for detecting local defects on a reflective surface (13), wherein the device can produce at least one pattern (12) for reflection on the reflective surface and comprises at least one camera (1) and a data processing unit (7), wherein the pattern comprises at least one essentially line-shaped light-dark transition, wherein the positioning and orientation of the camera are known, wherein the camera is configured to record the pattern reflected on the surface, to record the pattern reflected on the surface and to generate image data of the reflected pattern and to transmit them to the data processing unit, wherein the data processing unit is configured to determine local defects on the surface based on an evaluation of at least one property of the at least one light-dark transition in the image data of the reflected pattern, characterized in that the at least one property comprises a contrast in a predetermined area of the at least one light-dark transition in the image data.

7. The device according to claim 6, characterized in that further properties of the at least one light-dark transition in the image data comprise a shape of the at least one light-dark transition in the image data and / or a change in brightness over a predetermined distance across the at least one light-dark transition in the image data.

8. The device according to any one of claims 6 to 7, characterized in that the position of the pattern with respect to the surface is unknown.

9. The device according to any one of claims 6 to 8, characterized in that the pattern comprises at least one stripe forming two substantially line-shaped light-dark transitions, and the data processing unit is configured to evaluate the width of the at least one stripe to determine local defects on the surface.

10. The device according to any one of claims 6 to 9, characterized in that the pattern and / or the surface are configured such that they perform a relative movement with respect to each other during the recording of the reflected pattern by the camera.

11. A computer program product comprising instructions effecting that the device of any one of the claims 6 to 10 performs the corresponding steps of the method according to any one of claims 1 to 5.

12. A computer readable medium storing the computer program product according to claim 11.

Images